Acute respiratory distress syndrome: Rethinking and research priorities

Substantial progress has been made in the treatment of acute respiratory distress syndrome (ARDS), from its first description to the most recent updates in its definition and management guidelines.^[1,2] However, significant uncertainties persist regarding the optimal therapeutic strategies. These gaps highlight the need for a reevaluation of clinical practices and the initiation of innovative research to address the resolved challenges.

Oxygen targets of ARDS: High or low? The primary objective of oxygen therapy for patients with ARDS is to maintain adequate oxygenation. The National Heart, Lung, and Blood Institute ARDS Clinical Trials Network recommends a target partial pressure of arterial oxygen (PaO₂) between 55 mmHg and 80 mmHg, although supporting evidence remains limited. In the LOCO2 study, patients receiving conservative oxygen therapy (target PaO₂: 55–70 mmHg; pulse oxygen saturation [SpO₂]: 88–92%) were at greater risks of both mesenteric ischemia and 90-day mortality (44.4% vs. 30.4%) than those receiving liberal oxygen therapy were (target PaO₂, 90–105 mmHg; SpO₂ ≥96%).^[3] However, the ICU-ROX and HOT-ICU trials revealed no statistically significant differences in primary outcomes between liberal and conservative oxygen strategies.^[4,5] Conversely, the Oxygen-ICU and HyperS2S trials revealed that the risk of mortality increased as oxygen targets increased.^[6,7] These conflicting results underscore the ongoing uncertainty surrounding optimal oxygenation targets for ARDS patients.

Given the heterogeneous nature of ARDS, the impact of oxygen therapy likely varies on the basis of individual patient characteristics. Using a machine learning model, Buell et al^[8] demonstrated that the predictive effect of oxygenation targets on 28-day mortality varied widely, from a 27.2% absolute reduction to a 34.4% increase, depending on factors such as age, sex, comorbidities, and the presence of brain injury or cardiovascular disease. The findings of the study emphasize the need for personalized oxygen therapy based on patient-specific factors.

Currently, there is a significant gap in clinical practice, as no existing approaches are available to guide the personalization of oxygen therapy for patients with ARDS. Consequently, future research should focus on developing models to assess the effect of individualized oxygen therapy for ARDS patients and establish personalized oxygen therapy strategies using machine learning. Specifically, which ARDS patients benefit from higher oxygen targets, and which patients may benefit from more conservative oxygen strategies? Furthermore, randomized controlled trials (RCTs) are essential for evaluating the impact of personalized oxygen therapy strategies on clinical outcomes.

Noninvasive oxygenation strategies: What is the optimal approach? Optimal noninvasive oxygenation strategies can prevent intubation and improve outcomes in ARDS patients. Compared with standard oxygen therapy, high-flow nasal oxygen (HFNO) therapy, helmet or face mask noninvasive ventilation (NIV), and continuous positive airway pressure (CPAP) therapy have been shown to reduce the need for endotracheal intubation in adult patients with acute hypoxemic respiratory failure (AHRF). However, the ideal noninvasive strategy remains uncertain. A randomized trial comparing HFNO therapy with standard oxygen therapy and NIV revealed no effect on the overall intubation rate. However, HFNO therapy reduced the intubation rate in patients with PaO₂/FiO₂ ≤200 mmHg.^[9] The inferior performance of NIV relative to that of HFNO therapy may be explained by the increased transpulmonary pressure swings, increased tidal volume, and consequent increased risk of patient self-inflicted lung injury (P-SILI),^[10] which is associated with treatment failure and death.

Compared with HFNO therapy, CPAP therapy, which maintains constant positive pressure during both inspiration and expiration, generates a greater positive end-expiratory pressure (PEEP), thereby facilitating alveolar recruitment. Furthermore, CPAP therapy limits transpulmonary pressure swings and tidal volume, thereby reducing the risk of P-SILI. Physiological studies revealed that compared with HFNO therapy, CPAP therapy may offer superior oxygenation, with similar inspiratory effort, breathing frequency, and patient comfort.^[11] In a recent randomized trial involving 85 patients with AHRF, CPAP therapy was found to be more effective than HFNO therapy in decreasing the risk of meeting the predefined intubation criteria.^[12] A network meta-analysis suggested that CPAP therapy may be superior to both HFNO and NIV in reducing mortality and intubation rates.^[13]

Therefore, large-scale, prospective RCTs are needed to determine whether CPAP therapy is more effective than HFNO therapy or NIV in treating ARDS and to identify which patient subgroups are most likely to benefit from each treatment approach. In addition, physiological studies are also needed to compare the effects of these strategies on lung ventilation-perfusion matching.

Lung-protective ventilation: standardized or personalized? According to international guidelines, lung-protective ventilation strategies involving low tidal volume and PEEP titration are recommended to improve gas exchange and reduce the risk of ventilator-induced lung injury in ARDS patients. However, the effectiveness of these strategies may differ due to the heterogeneity of ARDS. Previous studies revealed that ventilation with a lower tidal volume reduces the risk of mortality, primarily in patients with high respiratory system elastance.^[14] A higher PEEP may be more beneficial for patients with nonfocal lung morphology.^[15] A personalized mechanical ventilation approach based on lung physiology or morphology may improve clinical outcomes.

A multicenter RCT comparing personalized mechanical ventilation with standard of care revealed no difference in 90-day mortality rates between the groups. However, when excluding 21% of misclassified patients in the per-protocol analysis, personalized ventilation demonstrated a significantly lower mortality rate.^[16] This indicates that accurate patient phenotyping is a prerequisite for implementing personalized mechanical ventilation. ARDS subphenotypes can guide the selection of the most appropriate PEEP. A secondary analysis of the ALVEOLI trial revealed that a higher PEEP was more effective in patients with a hyperinflammatory phenotype.^[17] Using data collected hourly during the first 4 days of invasive ventilation, Chen et al^[18] identified three distinct ARDS longitudinal phenotypes: Class 1 had fewer abnormal values, less organ dysfunction, and the lowest 28-day mortality; Class 2 was characterized by pulmonary mechanical dysfunction; and Class 3 was correlated with extrapulmonary dysfunction and had the highest 28-day mortality rate. The phenotype changed at least once during the first 4 days of mechanical ventilation in 56.9% of patients. Considering the heterogeneity treatment effect, patients with Class 2 ARDS can benefit from a lower PEEP, whereas those with Class 3 benefit from a higher PEEP.

The potential benefits of personalized ventilation are evident. However, significant limitations remain owing to the lack of refined lung morphology classification methods and the absence of integrated techniques, such as lung ultrasound, for real-time bedside guidance. Moreover, the optimal ventilation strategy for different ARDS subphenotypes, particularly from a physiological standpoint, remains inadequately explored. Although phenotyping can potentially guide the individualization of mechanical ventilation strategies, the dynamic nature of changes in phenotype over time underscores the need for further investigation into how these evolving profiles should inform ongoing treatment decisions.

Awake-prone position: is a longer duration better? Awake-prone positioning (APP) has been shown to improve oxygenation and reduce intubation rates in coronavirus disease 2019 (COVID-19) patients with AHRF, with evidence of a dose-response relationship. Previous observational studies and meta-analyses suggest that longer durations of daily APP are associated with lower risks of intubation and death.^[19,20] In a multicenter RCT comparing prolonged APP (12 hours/day) with shorter APP (5 hours/day), Liu et al^[21] reported that, compared with standard care, prolonged APP significantly reduced the intubation rate within 28 days.

Considerable challenges remain regarding the implementation of the APP. The optimal duration of APP remains uncertain; A secondary analysis of the prolonged APP trial revealed that 8 to 12 hours per day of APP may be the most beneficial duration, extending prone positioning beyond this timeframe did not yield additional clinical benefits (unpublished) and the generalizability of these findings is limited; Thus further validation is needed in prospective studies. Moreover, variations in the course of disease and clinical characteristics may also influence the efficacy of APP, and future research should focus on identifying which patients would benefit the most. Finally, the utility of APP in non-COVID-19 ARDS patients requires further exploration.

In conclusion, the management of ARDS is complicated by its heterogeneous nature, thus personalized therapeutic approaches are needed. Future studies should focus on developing evidence-based, individualized treatment protocols that consider patient-specific characteristics, subphenotypes, and dynamic responses to therapy.

Conflicts of interest

None.

Funding

This work was supported by grants from the Noncommunicable Chronic Diseases-National Science and Technology Major Project (No. 2023ZD0506500) and the National Natural Science Foundation (Nos. 82341032, 82402565, and 81930058).